AI is narrowing the discovery funnel to three results, and most brands are not ready
Where a search engine returns ten links, an AI assistant returns one to three. That compression changes what it means to compete for attention, restructures how creative is produced, and is already forcing organizational adaptation across the marketing industry.
The discovery funnel is narrowing in ways that most brand and marketing teams have not yet priced in. David Conforti puts the stakes plainly: where a Google search returns ten links, an AI assistant returns one to three results. “If you’re not in the top three, you are irrelevant. You might as well be 100.” That compression does not just change SEO strategy. It changes what it means to compete for attention at all. Ranking fourth in an AI-mediated query is functionally the same as not ranking.
The production side of advertising is already absorbing the impact. Cliff Weitzman reports that Speechify now tests close to 1,000 AI-generated ads per day, alongside roughly 8,000 human-made creatives produced monthly. That ratio, nearly a thousand machine-generated variants daily against a large but finite human output, is not a pilot program. It is a production architecture. The implication is that creative volume is decoupling from headcount in much the same way that research output has decoupled from team size in other domains.
What Weitzman’s numbers describe at the production level, Liva Ralaivola describes at the user-experience level for the near future. Within two to three years, Ralaivola argues, real-time generative ad creative generation will become feasible as a standard capability. Within five years, the interaction model itself will shift: users will ask AI assistants to surface a small, curated set of product options and will actively consent to advertising as part of that exchange. The query becomes the permission. That is a meaningful departure from the current paradigm, where advertising is something served to users rather than something users solicit.
If you're not in the top three, you are irrelevant. You might as well be 100. David Conforti
Martin Casado points to a capability that makes Ralaivola’s five-year scenario feel less speculative than it might otherwise seem. Image-based product queries, the kind where a user holds up a photo of a coat and asks what it is and where to buy it, already work. Casado notes that the same query would not have functioned reliably five years ago and certainly not ten. The infrastructure for AI-curated product discovery is not being built from scratch. It is being extended from a foundation that is already operational.
The organizational side of the transformation is where Rory O’Driscoll’s analysis carries the most weight. He estimates that today roughly 10% of CMOs can run AI-driven campaigns using agents, and argues that competitive pressure will push that figure to 40% over time, not because of enthusiasm for new tools but because those who cannot adapt will be forced out. That is a Darwinian framing, and it is probably the right one. The adoption curve here is not driven primarily by curiosity. It is driven by the cost differential between teams that can run a thousand creative variants daily and those that cannot.
Taken together, the picture that emerges is of a transition happening simultaneously at three levels: how creative is produced, how products are discovered, and how organizations are structured to compete. Each of those levels is moving on a slightly different timeline, but they are moving in the same direction. Brands that treat AI-mediated discovery as a future concern rather than a present one are already behind Speechify’s daily testing cadence. The brands that treat it as someone else’s concern entirely are the ones Conforti’s math renders invisible.